Disclaimer: The views and opinions expressed in this blog post are Darryl Chiang’s and do not reflect those of his employer, Google. He is sharing his personal perspective and is not speaking on behalf of the company.
In the second edition of our #miaME series, Darryl Chiang explores the transformative potential of AI-driven contract standardization in the legal industry. He highlights the tension between AI's current capabilities and the goal of harmonizing contracts, discussing the challenges and opportunities AI presents for law firms, in-house teams, and their clients.
Darryl also addresses ethical and operational considerations as AI becomes more integrated into contract negotiations, offering a thought-provoking perspective on the future of legal operations.
Until his Keynote on Day 2, please enjoy his interview!
How can AI-driven contract standardization reshape the relationship between law firms, in-house legal teams, and their clients, especially in balancing efficiency with personalized legal service?
I think there’s a tension between what AI can currently do versus the long-term direction that many of us believe the profession should actually go in. Today, AI can parse enormous numbers of badly-drafted, needlessly bespoke contracts and deliver to human lawyers a redline with comments at the level of a good summer associate. In the long-term, however, it would be better for everyone to use harmonized, standardized agreements for most transactions, resulting in less effort for both humans and machines. Existing AI solutions create an almost false sense of efficiency because the current efficiency gains perversely encourage companies to keep churning out long, turgid, custom agreement templates–and then let AI bots engage in a battle of the forms. Instead, we need to move towards a true state of efficiency: where 80% of standard transactions are handled by consensus-driven, open-source contracting terms like oneNDA and oneSaaS.
How can AI help us get to that ideal state? First, AI's powerful data analytics capabilities can show us the hidden consensus behind millions of differently-worded clauses that actually mean the same thing. Meanwhile, AI's creative writing skills can also instantaneously convert legalese into human-readable, plain-language wording. Combining these capabilities, open-source contract standardization organizations like oneNDA could use AI to review thousands of versions of a force majeure clause (for example); collate them into a “consensus” version that represents what all these clauses have in common; include a few “optional” add-on sub-clauses that a meaningful minority of the variants contain; and translate the wording into plain language. Relying on AI to do this initial work–based on a data-driven approach and without human “pride of authorship”--would not only be a better use of AI contract review resources, but could also help accelerate humans’ adoption of the standardized results.
Moving to this future state will not only be theoretically desirable, but will also be technically and environmentally imperative. We need to reduce the massive energy consumption, machine costs, and environmental degradation that will follow. By moving to contract standardization, we can drastically reduce the data our models have to plow through.
And even after AI plows through wordy contracts today, humans still end up in the loop for nearly every transaction, laboriously reading AI’s comments and manually accepting or rejecting AI’s proposed changes. If 80% of our transactions are instead handled by standardized terms, in-house counsel, firms, and their clients will have time to focus on the small percentage of terms that are truly novel or worth negotiating–and that really warrant humans’ time. Bringing the stakeholders together for only these truly consequential discussions will potentially forge stronger bonds among people because the time they spend together will be strategic and meaningful.I think there’s a tension between what AI can currently do versus the long-term direction that many of us believe the profession should actually go in. Today, AI can parse enormous numbers of badly-drafted, needlessly bespoke contracts and deliver to human lawyers a redline with comments at the level of a good summer associate. In the long-term, however, it would be better for everyone to use harmonized, standardized agreements for most transactions, resulting in less effort for both humans and machines. Existing AI solutions create an almost false sense of efficiency because the current efficiency gains perversely encourage companies to keep churning out long, turgid, custom agreement templates–and then let AI bots engage in a battle of the forms. Instead, we need to move towards a true state of efficiency: where 80% of standard transactions are handled by consensus-driven, open-source contracting terms like oneNDA and oneSaaS.
How can AI help us get to that ideal state? First, AI's powerful data analytics capabilities can show us the hidden consensus behind millions of differently-worded clauses that actually mean the same thing. Meanwhile, AI's creative writing skills can also instantaneously convert legalese into human-readable, plain-language wording. Combining these capabilities, open-source contract standardization organizations like oneNDA could use AI to review thousands of versions of a force majeure clause (for example); collate them into a “consensus” version that represents what all these clauses have in common; include a few “optional” add-on sub-clauses that a meaningful minority of the variants contain; and translate the wording into plain language. Relying on AI to do this initial work–based on a data-driven approach and without human “pride of authorship”--would not only be a better use of AI contract review resources, but could also help accelerate humans’ adoption of the standardized results.
Moving to this future state will not only be theoretically desirable, but will also be technically and environmentally imperative. We need to reduce the massive energy consumption, machine costs, and environmental degradation that will follow. By moving to contract standardization, we can drastically reduce the data our models have to plow through.
And even after AI plows through wordy contracts today, humans still end up in the loop for nearly every transaction, laboriously reading AI’s comments and manually accepting or rejecting AI’s proposed changes. If 80% of our transactions are instead handled by standardized terms, in-house counsel, firms, and their clients will have time to focus on the small percentage of terms that are truly novel or worth negotiating–and that really warrant humans’ time. Bringing the stakeholders together for only these truly consequential discussions will potentially forge stronger bonds among people because the time they spend together will be strategic and meaningful.
As law firms and in-house teams increasingly use AI in contract negotiation, what ethical, operational, and strategic considerations are essential to maintain client trust and accountability?
While we’ve all seen egregious cases in the news of lawyers using general-purpose AI chatbots to write briefs for them–only to discover that they’re full of hallucinatory case citations and sophistic arguments–the error rate is much lower when commercial lawyers use purpose-built AI tools that are fine-tuned and grounded on the company or firm’s own contract templates, fallbacks, and contracting guidelines.
Of course, ethically, current AI tools are still like summer associates whose work needs to be reviewed and checked by licensed lawyers before it goes out the door. Operationally, there may be concerns that some tech companies are actually training on or surreptitiously using the privileged and confidential documents we upload to their AI systems. And strategically, lawyers want to maintain a direct relationship with our clients rather than being disintermediated by AI tools.
In the short term, bar associations are scrambling to clarify ethical rules, tech companies are developing additional safeguards and certifications to allay enterprise customers’ concerns, regulators are working on guardrails, and legal professionals are still emphasizing the need for their human-in-the-loop judgment to double check the results of this still-nascent technology. But I think these issues are relatively contained to the extent AI is still very limited in its capabilities.
The greater challenges will come in the next generation of AI. AI is already capable of outperforming humans in specific tasks–including passing the bar exam–but tech companies are aiming for “artificial general intelligence” (AGI), loosely defined as AI that matches or surpasses humans across a wide range of cognitive tasks. And artificial “superintelligence” will massively exceed humans’ thinking abilities. At that point, it may be hard for us to even comprehend what AI is independently doing, or to demand that humans-in-the-loop review and approve AI’s work. As a profession, we need to quickly help define the legal prerequisites for safe AGI that’s aligned with humans and our larger ecosystem. Ideally, AGI could then provide humans with an independent and compelling source of superintelligent analysis that focuses not just on what we’ve been doing, or what we currently do, but on what we ought to do–whether that’s agreeing to fair and efficient contract terms, resolving individual, local, and global conflicts, or solving our energy and climate crises.
With legal operations leveraging AI tools, how can in-house teams and law firms ensure that AI adoption strengthens, rather than diminishes, their role as strategic business partners within their organizations?
As impressive as our current AI contracting tools are, they’re still like advanced spell check and grammar check–they methodically flag potential issues and propose logical solutions, but humans uniquely possess the judgment to know which of those flags are important and which aren’t. Humans are also the only ones that can express the nuanced tone and syntax required to navigate contentious, illogical, emotionally-charged negotiations. In the short- and medium-term, I think both in-house lawyers and law firms can use AI tools to speed up the first few steps of contract review. This may ultimately strengthen the lawyers’ role as strategic business partners because it leaves more time for the lawyers to zero in on the small percentage of truly critical issues to discuss with their clients, and to focus those discussions on higher-level principles and strategic implications.
What are you most looking forward to hearing or experiencing at Future Contracts Miami?
I’m most looking forward to having a dialogue–rather than a one-way discussion of–the future of contracts. I was originally invited to deliver a keynote speech for the conference, but I think that having a conversation with as many participants as I can would be much more interesting and exciting. We’re only at the beginning stages of the AI revolution, so I’m looking forward to learning from others more than sharing my current perspectives–particularly about the prospect of adopting standardized, open-source terms and templates that are written in plain language.
With the event set in Miami, there’s so much to explore! From food and unique experiences to stunning sights, is there a particular spot or activity in Miami that you’re especially excited about?
I’m very interested in architecture, so I’m especially excited to take an Art Deco walking tour. I also confess that I’m a 1980s pop culture fan, and would love to take a Miami Vice tour!
Future Contracts Miami is thrilled to explore groundbreaking tools like ContractCrab and more that are transforming the legal landscape. Be part of #FLMiami2025 to gain cutting-edge insights.
